In real/demo account losing, while back tests start on 100$ and gives Millions/Trillions with DD about 5% - page 2

 
Sergey Golubev:

Those EAs should be fixed for real/demo trading.

example - this thread - Help me to find the bug in my "holy grail" (and the EA is posted on the first post of the thread as free to download by source code).


fix what, Sergey

 
EnriqueD:

Not sure how you come up with 99% the screenshots show 30% and 78%.

It doesn't even matter.

In fact the closest to real results is on bar open times only, because then you excrete the ticks that are responsible for these crazy numbers. 

The best test will be on M1 and bar open times only, in this case every M1 bar will act as a real market tick, in stead of a series of induced or injected artificial ticks between high and low (which are causing the problem here)

 

This is 99.9% 

but shows N/A I don't know why


.

I think the modelling quality is not a matter in this question as Marco says

 

And there is the thread in Chinese part of the forum with such "the tools" (discussion, opinions, fixing to become a normal one).

Those are the comments from that Chinese thread:

  • "To interpret this graph - take a look at the test history from the start and to the end time of testing, and look at this delightful amplitude/time line."


  • "Since 1971? My sister was born at this time, now she is almost 50 years old. Since foreign exchange transactions do not allow you to make quick money, there's nothing to talk about."


And I like this comment from the Chinese thread more and more:

  • "I look at these charts, and I feel better and better."

 
Mohammad Soubra:

the 1st one is a paid highest quality


78% still means 22% thin air, this means, the results are not to be trusted.

 
EnriqueD:

78% still means 22% thin air, this means, the results are not to be trusted.

The results are never to be trusted.

It makes me laugh seeing people advertise their 'Million Dollar Backtests' hahaha

 
Marco vd Heijden:

The results are never to be trusted.

It makes me laugh seeing people advertise their 'Million Dollar Backtests' hahaha


No, I believe that one day someone will do that in real money

maybe me


 

People already do that in real money Mohammad, but if you run their algo in the backtester it will fail miserably.

And many times their algo can not even be tested because they use multiple symbols or other techniques...

I meant to say that backtest results are never to be trusted.

 
Mohammad Soubra:

fix what, Sergey


If you fix it, and backtest it one more time after fixing so look at the backtesting graph: if you feel better and better looking on this graph so fix your EA once again (suggestion from the Chinese part of the forum).
backtesting therapy

 

Here, a question has been came in my mind. while I tried to ask it, why most people are not trusting the back test data

let's say the paid data, it is from a Swiss Bank, is it fake also?? are they thieves ??! no they are not ... but let us discuss the subject while you are believing the Bach Tester